(CANCER RESEARCH 58. 5123-5129. November
`
`15. 19981
`
`NH2-terminally Truncated HER-2/neu Protein: Relationship with Shedding of the
`Extracellular Domain and with Prognostic Factors in Breast Cancer1
`
`Tracy A. Christianson,2
`and Gail M. Clinton4
`
`Joni K. Doherty,2 Yuzhong J. Lin,2J Elizabeth E. Ramsey, Ron Holmes, Edward J. Keenan,
`
`of Biochemistry' anil Molecular Biology fT. A. C., Y. J. L., G. M. C.j. Cell and Developmernal Biology // K. D.}, and Physiology and Pharmacology
`Departments
`E. J. K. I, Oregon Health Sciences University. Portland, Oregon 97201
`
`/£.£./?.. K. H..
`
`ABSTRACT
`
`We identified an MI ,-U>rmm;ill> truncated HER-2/neu product of M,
`95,000 with in vitro kinase activity by Western blotting and immunopre-
`cipitations using domain-specific
`antibodies. p95 levels correlated with the
`extracellular
`domain (ECD)
`shed from different
`cells under varied con
`ditions. Both ECD and p95 were at
`-20-fold
`lower levels
`in SKOV3
`ovarian carcinoma
`cells, as compared to BT474 breast carcinoma
`cells.
`Both were stimulated by treatment of cells with the phorbol ester tumor
`promoter phorbol 12-myristate
`13-acetate and the lysosomotrophic
`agent
`chloroquine. The hydroxamate
`inhibitor of metalloproteases,
`TAPI, sup
`pressed both p95 and ECD in a dose-dependent
`fashion, with maximal
`inhibition at SlO /<M in BT474 cells. Cancer
`tissues were analyzed by
`Western blotting and scored for p95HER-2/neu
`and for pl85HER-2/neu
`expression. Breast and ovarian cancer tissues were both found to express
`p95HER-2/neu
`in addition to pl85HER-2/neu.
`Of 161 breast cancer tis
`sues, 22.4% expressed p95, 21.7% overexpressed
`pl85, and 14.3% were
`p95 positive and overexpressed p 185. A higher proportion of node-positive
`patients
`(23 of 78) than node-negative
`patients
`(9 of 63) expressed p95 in
`all tumors combined (P = 0.032).
`In the group that overexpressed
`pl85,
`those that contained p95 were associated with node-positive patients (15 of
`21), whereas
`those that were p95 negative were associated with node-
`negative patients (8 of 11; P = 0.017). Neither p95- nor pl85-rich patients
`significantly
`correlated with tumor size or with hormone
`receptor status
`in this study. Our findings
`show that breast cancers, which express
`the
`HER-2/neu oncogene, are heterogeneous with respect
`to HER-2/neu pro
`tein products. p95HER-2/neu
`appears
`to distinguish
`tumors
`that have
`metastasized to the lymph nodes from those in node-negative
`patients.
`
`INTRODUCTION
`The HER-2/neu (erbB-2) gene encodes a RTK.5 which is a member
`of the EGF receptor
`family ( 1). Overexpression
`of HER-2/neu has
`been observed in tumors arising at many sites,
`including non-small
`cell
`lung (2), colon (3), prostate (4), ovary, and breast (5). In human
`breast cancer,
`in which HER-2/neu involvement has been extensively
`studied, overexpression
`occurs in 15-30% of the cases (see Ref. 6)
`and predicts a significantly lower survival
`rate and a shorter time to
`relapse in patients with lymph node-positive disease (5-8). The sig
`nificance of HER-2/neu in node-negative patients is controversial,
`and
`thus far, its clinical utility as a prognostic indicator
`is limited (8, 9).
`Various approaches are being taken toward HER-2/neu-targeted
`ther-
`
`Received 7/7/98; accepted 9/15/98.
`The costs of publication of this article were defrayed in part hy the payment of page
`charges. This article must
`therefore he hereby marked advertisement
`in accordance with
`18 U.S.C. Section 1734 solely to indicate this fact.
`1Supported by National Cancer
`Institute Grant CA-71447 and Department of Defense
`Breast Cancer Research Program Grant DAMD17-6204.
`J. K. D. is a predoctoral
`fellow
`of the Department of Defense Breast Cancer Research Program.
`- The first three authors contributed equally to this work.
`* Present
`address: Cascade Biologies.
`Inc.. 4475 Southwest
`
`Scholls Ferry Road.
`
`Portland. OR 97225.
`4 To whom requests for reprints should be addressed, at Department of Biochemistry
`and Molecular Biology. Oregon Health Sciences University. 3181 Southwest Sam Jackson
`Park Road. Portland. OR 97201. E-mail: clinton@ohsu.edu.
`kinase; ECD. extracellular
`5 The abbreviations
`used are: RTK.
`receptor
`tyrosine
`domain: EGF. epidermal growth factor; HMEC. human mammary epithelial cell: TBS.
`Tris-buffered
`saline: ER. estrogen receptor; PR. progesterone
`receptor: PMA. phorbol
`12-myristate
`13-acetate: TAPI.
`/V-{D.L-|2-(hydroxyaminocarbonyl)methyl|-4-methylpen-
`tanoyl)
`i.-3-(2'naphthyl)-alanyl-i.-alanine,2-aminoethylamide.
`
`apeutics many of which are based on antibodies that are specific to the
`ECD of the transmembrane
`protein, which either down-regulate
`re
`ceptor
`function or target recombinant
`toxins, with the goal of specif
`ically killing HER-2/neu-expressing
`tumor cells (8-10).
`In addition to the full-length transmembrane
`product of the HER-
`2/neu gene, p 185. a truncated product corresponding
`to the ECD is
`released from breast carcinoma cells in culture by regulated proteol-
`ysis ( 11-13) and is also produced from an alternative transcript
`( 14).
`HER-2/neu ECD is elevated in the serum of patients with breast (15),
`ovarian (16), and prostate
`cancer
`(17). Several
`studies of breast
`cancers estimate that s69f> of patients with early-stage disease, ~25%
`of patients with metastatic and locally advanced disease, and >50% of
`patients with recurrent metastatic disease have elevated serum ECD
`(see Ref. 18). Elevated ECD in serum is associated with overexpres
`sion of HER-2/neu in tumor tissue and also reflects tumor
`load (19,
`20). Soluble HER-2/neu is a marker of metastatic disease and may
`predict
`recurrence (19), shortened survival
`(20-23),
`and response to
`antiestrogen therapy in advanced-stage
`patients (24. 25). Serum ECD
`has also been reported to neutralize
`the activity of anti-HER-2/neu
`antibodies
`targeted to the ECD (26, 27). possibly allowing escape of
`HER-2-rich tumors from immunologie-ai control.
`an NH2-
`to create
`Proteolytic
`release of
`the ECD is expected
`terminally truncated, membrane-associated
`fragment with kinase ac
`tivity. Cellular
`fragments created by ectodomain shedding have been
`described for the colony-stimulating
`factor
`receptor
`(28),
`the TrkA
`neurotrophin receptor (29), the Axl receptor (30), and HER-4 (31), but
`a truncated cellular product of HER-2/neu shedding has not yet been
`identified. The truncated colony-stimulating
`factor receptor was found
`to have in vitro kinase activity (28), and the cytoplasmic HER-4,
`induced by phorbol ester
`tumor promoters,
`had little or no kinase
`activity (31 ), whereas a truncated HER-4 found in cells treated with a
`proteosome
`inhibitor was an active kinase (32).
`full-length
`Several
`lines of evidence
`indicate
`that
`the ECD of
`transmembrane
`receptors exerts a negative regulatory constraint on
`their signaling activity. The engineered deletion of a region of the
`HER-2 ECD was found to enhance its oncogenic potency (33.-36).
`This has also been illustrated by engineered removal of the ECD from
`the EGF receptor and by the oncogenic
`potency of viral encoded
`v-erbB, v-kit. and v-ros, that are missing regions of the ECD found in
`their normal cellular counterparts
`(37). Naturally occurring mutant
`EGF receptors with NH2-terminal
`truncations have been identified in
`several human carcinomas
`(38) and have constitutive
`signaling activ
`ity and enhanced oncogenic transforming activity in cell culture and
`animal models (39, 40).
`the NH2-terminally
`Here, we sought
`to identify and characterize
`truncated HER-2/neu protein and examine its correlation with ECD
`shedding and association with breast cancer pathological
`factors.
`
`MATERIALS AND METHODS
`
`lines were obtained from the American Type
`Cell
`Cells and Antibodies.
`Culture Collection
`(Manassas. VA), except
`the 3T3 cells
`transfected with
`HER-2/neu cDNA (17-3-1: provided by Applied BioTechnology.
`Inc.. Cam
`bridge. MA), and the HMECs (provided by Dr. Gary Shipley. OHSU. Portland,
`
`5123
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on November 12, 2014. © 1998 American Association for Cancer
`Research.
`
`IMMUNOGEN 2048, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TRUNCATED HER-2/neu KINASE AND BREAST CANCER
`
`found to overexpress HER-2/neu in numerous clinical
`30% of breast cancers
`studies (5-10).
`In the samples that had detectable p95, its level ranged from 10
`to 100% ofthat of pi 85. In this pilot study, specimens were scored as positive
`if p95 was detected at a a 10% proportion of p 185 by 2 h of exposure of the
`membrane
`to film. Because of the high titer of the primary antibody,
`anti-
`neu(C),
`there were rarely any background bands, even when the immunoblots
`were exposed to film for 2 h.
`HER-2/neu Tissue Extract ELISA. Aliquots of membrane-rich fractions
`prepared from breast cancer
`tissue as described above were assayed using the
`Triton Diagnostics
`c-erbB-2 Tissue Extract EIA kit (Ciba Corning) according
`to manufacturer's
`instructions. This assay uses two monoclonal
`antibodies
`
`the HER-2/neu ECD. The HER-2/neu units per milligram of protein in
`against
`the specimens were calculated from a calibration curve generated by plotting
`the HER-2/neu concentration
`of the calibration
`standards versus the absorb-
`ance obtained from the immunoassay.
`Clinical Data. This
`investigation of human tissues was reviewed and
`approved by the Institutional Review Board Committee on Human Research.
`A computer database contains clinical
`information
`on each patient, coded to
`protect
`the individual's
`identity,
`and includes
`age, nodal status,
`size of the
`
`primary tumor, age of the patient, stage of disease at diagnosis, ER levels, and
`PR receptor
`levels. Specimens were considered ER positive and PR positive if
`they contained
`at
`least 10 fmol of specific binding sites per milligram of
`cytosolic proteins. The stages of the specimens were as follows: 1 stage 0, 32
`stage
`I, 56 stage
`II, 45 stage
`III, and 13 stage
`IV specimens.
`Fourteen
`specimens were of unknown stage. The average age of the patients was 60
`years. The eight ovarian cancer
`tissues included three that were grade III and
`five that were grade IV.
`
`RESULTS
`
`Identification of NH2-terminaIly Truncated HER-2/neu Protein
`with Kinase Activity. 3T3 cells transfected with HER-2/neu cDNA
`(17-3-1 cells)
`release
`soluble ECD by proteolytic
`processing
`of
`pl85HER-2/neu
`(12). To detect
`truncated cytoplasmic products, we
`resolved 17-3-1 extracts in gels and immunoblotted with antibodies
`against
`the COOH terminus of the HER-2/neu product
`[anti-neu(C)].
`Two major protein products were detected in cell extracts:
`the full-
`length pl85HER-2/neu
`and a truncated protein of M, -95,000
`(Fig. 1,
`Lane I). Extracts were immunoprecipitated,
`and the MT95,000 protein
`as well as pl85HER-2/neu were phosphorylated
`in the immune com
`plex with [-y-32P]ATP (Fig. 1, Lane 2). A monoclonal
`antibody that
`was specific for the NH2-terminal
`region of pl85HER-2/neu
`[anti-
`
`OR) were cultured from tissue obtained from reduction mammoplasty. Anti-
`peptide antibody against
`the COOH terminus of pl85HER-2/neu,
`anti-neu(C),
`has been described previously (41). Monoclonal
`antibody against
`the ECD of
`HER-2/neu was prepared
`as described
`(42) and was provided by Applied
`BioTechnology.
`Inc.
`Cell Culture.
`17-3-1 cells were cultured in DMEM supplemented with 5%
`fetal bovine serum containing 0.4 mg/ml geneticin (G418; Life Technologies,
`Inc.). The human breast carcinoma
`cell
`line BT474 was cultured in RPMI
`medium supplemented with 10% FBS and 10 /ig/ml
`insulin. All other cell lines
`were grown in DMEM supplemented with I0"7r FBS and the antibiotic gen-
`lamicin at 0.05%.
`and Immune Complex Kinase Assays. Freshly
`Immunoprecipitations
`prepared cell
`lysates
`in TEDG buffer
`[50 mM Tris, 1.5 mM EDTA, 0.5 mM
`DTT, 10% glycerol
`(pH 7.5) with \% aprotinin. 2 mM phenylmethylsulfonyl
`fluoride, and 2 mM vanadatel containing
`1% NP40 were immunoprecipitated
`by incubation with antibody
`for 2 h with continuous
`shaking at 4°C as
`described (41). The immune complexes, bound to protein G-Sepharose
`(Phar
`macia), were washed twice with TEDG buffer and incubated 10 min on ice in
`a kinase reaction mixture containing 20 m.MHEPES (pH 8.0). 2 mM DTT. 25
`JIM vanadate,
`0.5% NP40,
`10 mM MnCK.
`1 ¡M ATP, and
`15 /¿Ciof
`(•y-''P]ATP(New England Nuclear). The immune complexes were washed
`
`three times with buffer, and the proteins were released by boiling for 2 min in
`SDS-polyacrylamide
`sample buffer.
`lysates or proteins from
`cell
`Western Blotting. Following SDS-PAGE,
`concentrated,
`conditioned medium were electroblotted
`onto nitrocellulose
`(Trans-Blot; Bio-Rad) using a semidry transfer unit (Bio-Rad) at 15 V for 20
`min per mini gel of 0.75-mm thickness
`(Mini-PROTEAN II electrophoresis
`cell: Bio-Rad) equilibrated with 25 mM Tris (pH 8.3), 192 mM glycine, 50 mM
`NaCl,
`and 20% methanol. Binding
`sites were blocked
`by incubating
`the
`membrane with 5% nonfat dry milk. After
`incubation with the primary anti
`body,
`the blot was washed twice for 15 min and four times for 5 min with TBS
`containing 0.05% Tween 20 and then incubated for 40 min with goat antirabbit
`or goat antimouse
`antibody conjugated
`to horseradish
`peroxidase
`(Bio-Rad)
`diluted in TBS-Tween. After incubation with secondary antibody,
`the blot was
`washed as described above with TBS-Tween
`and developed with chemilumi-
`nescent
`reagent
`(Pierce).
`tumor
`and Fractionation. About 0.1 g of
`Cancer Tissue Extraction
`tissue, which had been fresh-frozen and stored at —¿(cid:3)70°C,was minced on dry
`
`using a
`in TEDG buffer. Tissues were homogenized
`ice and suspended
`Brinkman polytron for 5-10-s bursts repeated two to three times with a chilled
`probe. Homogenates were centrifuged at 1500 x g for 10 min at 4°C.The lipid
`
`layer was removed with a wooden stick, and the supernatant was centrifuged
`for 20 min at 40,000 X g at 4°C.The lipid layer was collected with a wooden
`
`the supernatant was decanted, and the pellet containing the membranes
`stick,
`solubilized
`in TEDG buffer
`containing
`0.1% SDS for 20 min with
`was
`intermittent vortexing and clarified by centrifugation
`at 15,000 X g for 15 min.
`The protein concentration
`in the supernatant was determined
`by the Bio-Rad
`protein assay reagent, and aliquols were frozen at -80°C.
`
`of p95 and p 185 in Breast Cancer Tissues. Twenty ju.g of
`Analysis
`protein from the membrane
`traction prepared from each tumor sample were
`resolved under denaturing and reducing conditions by SDS-PAGE in 10% gels.
`Each gel also contained 3 /ng of protein from extracts of 17-3-1 cells to mark
`the migration of p 185 and p95 and to provide a standard for the entire study.
`Proteins were electrotransferred
`onto membranes
`as described above, which
`were incubated with anti-neu(C)
`diluted 1:10.000 in TBS-Tween
`20 at 4°C
`
`overnight with shaking and then incubated with a 1:10.000 dilution of goat
`antirubbit horseradish peroxidase-conjugated
`antibody (Bio-Rad)
`for 40 min at
`room temperature. To develop the blot,
`the membranes were incubated with
`chemiluminescent
`reagent
`(Pierce)
`for 5 min and then exposed
`to Kodak
`X-OMAT AR film for 1, 5, 20, and 120 min. To define the samples
`that
`overexpressed
`pl85HER-2/neu.
`specimens with HER-2 immunoassay
`values
`that were considered HER-2/neu-rich
`(400 units or greater)
`compared
`to
`samples with low HER-2/neu levels (<400 units) were characterized
`for their
`p 185 signal
`relative to the control
`17-3-1 cells by Western analysis. Those
`samples with a pi85 signal
`that could be detected by a 1-min exposure of the
`membrane
`to film and that was more than or equal
`to the pl85 level found in
`3 jig of 17-3-1 cells, as revealed by laser densitometric
`analysis of the film,
`were scored as highly positive. Using this method, we identified 21.7% of the
`samples
`that overexpressed
`pl85. This proportion
`is comparable
`to the 15-
`5124
`
`blot:
`
`a-neu(C)
`
`IP:
`o-neu
`(C)
`(N)
`
`(N) 1st
`(C)2nd
`
`200 kDa -
`
`116kDa-
`
`97 kDa
`
`a
`
`<pl85
`
`<p95
`
`Fig. 1. NH2-lerminally truncated HER-2/neu product with kinase activity. Twenty-five
`¿igof protein from 17-3-1 cells were Western blotted with anti-neu(C) diluted 1:10,000
`(Lane I).
`In Lanes 2-4, 400 /xg of protein were immunoprecipitated
`with anti-neu(C)
`(Lanes 2 and 4) or with monoclonal antibody against
`the ECD, anti-neu(N)
`(Lane 3), or
`were depleted of pl85HER-2/neu
`by extracting twice with anti-neu(N)
`and then immu
`noprecipitated with anti-neu(C)
`(Lane 4). The immune complexes were phosphorylated
`with [y-12P)ATP and analyzed by SDS-PAGE and autoradiography.
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on November 12, 2014. © 1998 American Association for Cancer
`Research.
`
`IMMUNOGEN 2048, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TRUNCATED HER-2/neu KINASE AND BREAST CANCER
`
`A
`TPA + Chlaroqulne
`
`Probe:
`
`a - neu (N)
`-
`+
`
`+
`
`the NH2-
`that
`indicating
`p95,
`immunoprecipitate
`neu(N)] did not
`terminal
`region was missing (Lane 3). To examine whether p95 had
`self-phosphorylating
`activity or was a substrate of
`the full-length
`RTK, pl85 was first removed from the cell
`lysate with anti-neu(N),
`and then p95 was immunoprecipitated
`with anti-neu(C).
`p95 was
`phosphorylated when p 185 levels were greatly depleted (Fig. 1, Lane
`4), indicating that it has kinase activity.
`p95 Kinase Activity Is in Human Breast Carcinoma Cells but
`not in Nontumorigenic Breast Epithelial Cells. The breast carci
`noma cell
`line, BT474, known to release
`soluble ECD (11) also
`contained
`two autophosphorylated
`HER-2/neu
`products, p 185 and
`p95, which were at elevated levels compared to the nontumorigenic
`breast epithelial cell
`line HBL-100 (Fig. 2). It was possible that p95
`could not be detected in the small amount of HBL-100 cells because
`they express low levels of HER-2 (43). To compensate
`for different
`levels of HER-2/neu expression
`(43),
`the amounts of extract
`from
`HBL-100, HMECs, and three breast carcinoma
`cell
`lines were ad
`justed, and proteins were phosphorylated with [y-^PJATP.
`p95 was
`detected in the low (MDA-MB-453)
`and high (BT474 and SKBR3)
`HER-2/neu-expressing
`breast carcinoma cells but not in the HBL-100
`nor HMEC cells, despite a robust signal from the HER-2/neu receptor,
`which migrated as a slightly smaller protein in the breast epithelial
`cells (Fig. 2).
`
`8 m
`
`p185
`
`p95
`
`mg cell protein
`
`0.2
`
`0.2
`
`2.0
`
`2.0
`
`Indicated amounts
`lines contain p95HER-2/neu.
`Fig. 2. Human breast carcinoma cell
`of cell lysates from BT474. HBL-100, MDA-MB-453. SKBR3. HMEC. and 17-3-1 cells
`were immunoprecipitated with anti-neu(C)
`and phosphorylaled
`as in Fig.
`I.
`
`IP:
`probe:
`
`anti-neu(C)
`anti-PTyr
`
`p185
`
`p95
`
`P S
`fraction of BT474
`p95 localized in the particulate
`Fig. 3. Tyrosine phosphorylated
`breast carcinoma cells. Particulate (Lane P} and soluble (Lane 5) fractions were prepared
`by incubation of IO7 cells in ice for 10 min in 3 ml of 10 mM Tris (pH 7.4). 10 niM NaCl.
`and 2 mM MgCU with 2 mM vanadate and protease inhibitors. Dounce homogenization,
`and centrifugation
`at 100.000 x g for 1 h. The pellet was resuspended
`in 3 ml of
`homogenization
`buffer, and 200 fig of protein from the paniculate
`fraction and an equal
`volume of the soluble fraction were immunoprecipitated with anti-neu(C) and analyzed as
`a Western blot with monoclonal
`antiphosphotyrosine
`antibody (Sigma Chemical Co.).
`
`5125
`
`pllO >
`
`B T
`
`PA + Chloroquine
`
`pl85
`
`p95 >
`
`SKOV3
`BT474
`Fig. 4. Expression of p95 and ECD in SKOV3 and BT474 cells. Cells were treated for
`24 h in serum-free medium with control vehicle or with 5(X) nvi phorbol ester PMA and
`50 /XMchloroquine. A. 5 ml of conditioned media were concentrated
`100-fold, denatured
`under nonreducing
`conditions,
`and aliquots normali/ed
`to cell extract protein were
`analyzed by Western blotting with anti-neu(N) monoclonal antibody at I fig/ml. B. 20 jig
`of cell proteins were analyzed by Western blotting using anti-neu(C). The results arc
`representative
`of three replicate experiments.
`
`and Is in the Membrane Frac
`p95 Is Tyrosine-phosphorylated
`tion from BT474 Cells. Tyrosine phosphorylation of tyrosine kinase
`receptors
`indicates their activation in signaling (9, 10). The tyrosine
`phosphorylation
`of p95 and its subcellular
`location were examined by
`fractionation
`of BT474 cell extracts
`into a soluble fraction and a
`particulate fraction, which were immunoprecipitated with anti-neu(C)
`and then subjected to Western blot analysis using monoclonal
`anti
`bodies against phosphotyrosine.
`Fig. 3 illustrates
`that tyrosine-phos-
`phorylated
`p95 fractionated with pi85
`in the particulate
`fraction,
`which contains the plasma membranes. p95 was further shown to be
`tyrosine-phosphorylated
`by first
`immunoprecipitating with antiphos-
`photyrosine
`antibodies and then probing the Western blot with anti-
`neu(C)
`(data not shown).
`p95 Corresponds
`to Levels of Soluble ECD Released from
`Different Cells. To examine the relationship of p95 to soluble ECD,
`their levels were compared in different cells under varied conditions.
`The basal
`levels of ECD and cellular p95HER-2/neu were first ex
`amined in two cell lines that overexpress HER-2/neu, BT474 and the
`ovarian carcinoma cell line SKOV-3, which was previously reported
`to produce low levels of ECD (13). The amount of p95 relative to
`p 185 and to cell protein was greatly elevated in BT474 cells, and
`correspondingly,
`the ECD in the extracellular medium from BT474
`cells, detected with anti-neu(N), was enhanced by > 10-fold compared
`to the SKOV3 cells (Fig. 4).
`is rapidly and transiently
`Shedding of several membrane proteins
`(44, 45). Although short-
`induced by phorbol ester tumor promoters
`term treatment with tumor promoters does not
`induce HER-2 shed
`ding (31 ), chronic administration of the phorbol ester PMA synergizes
`with chloroquine
`to stimulate release of soluble HER-2.6 To deter
`
`mine whether p95 and ECD were coordinately regulated, PMA (500
`nrn) and chloroquine (50 /XM)or the control vehicle were added to the
`culture media of BT474 and SKOV3 cells. At 24 h, the ECD levels in
`the extracellular media and p95 levels in the cell extract were ana
`lyzed. Soluble ECD was elevated
`severalfold
`in the conditioned
`
`s T. A. Christiansen. Y. J. Lin. and G. M. Clinton, unpublished observations.
`
`BT474 SKOV3
`Probe:
`a - neu (C)
`
`•¿(cid:3)§
`
`
`
`Downloaded from on November 12, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org
`
`
`Research.
`
`IMMUNOGEN 2048, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TRUNCATED HER-2/ncu KINASE AND BREAST CANCER
`
`Probe:
`
`a - neu (N)
`
`pllO>
`
`20 40
`10
`O 1
`TAP I (MM)
`
`B
`
`Probe;
`
`a-neu (C)
`
`O
`
`10
`
`1
`TAPI
`
`20
`(MM)
`
`40
`
`acid TAPI. BT474 cells in
`Fig. 5. p95 and ECD arc inhihiied by the hydroxamic
`serum-lree medium were treated lor 24 h with the control vehicle or with 1. 10. 20. and
`40 fiM TAPI (a gift from Immunex). A. the concentrated,
`conditioned media, normali/ed
`to the amount of cell extract, were analyzed by Western blotting with anli-neu(N). Similar
`results were obtained when 5 ^ig of protein from the conditioned media from each culture
`were analyzed. H. 20 ng of cell proteins were analyzed
`by Western blotting using
`anti-neu(C).
`
`medium from stimulated BT474 cells and SKOV3 cells, whereas p95
`was up-regulated
`~3-t'old in BT474 cells (Fig. 4). Overexposure
`of
`
`revealed that p95 in SKOV3 cell extracts was also
`the immunoblot
`stimulated about 3-fold by TPA and chloroquine
`(data not shown).
`A Metalloprotease
`Inhibitor Depresses Levels of p95 and ECD
`from BT474 Cells. The shedding of diverse transmembrane proteins
`has been found to be inhibited by hydroxamic acid-based compounds,
`which are potent metalloproteinasc
`inhibitors (46-48). We. therefore,
`tested the effects of different concentrations
`of the hydroxamic
`acid
`TAPI (47) on shedding of HER-2/neu ECD and on cell levels of p95.
`TAPI at 0-40
`jLtMwas added to cultured BT474 cells for 24 h, the
`ECD in concentrated
`conditioned media was analyzed by immuno-
`blotting with anti-neu(N).
`and p95 and p 185 were examined in cell
`extracts using anti-neu(C). The results in Fig. 5 show that production
`of ECD was partially inhibited at 1 /AMTAPI and maximally inhibited
`by 10 JIM TAPI. A residual amount of -10% of the ECD resisted
`inhibition by 40 /XMTAPI. The level of truncated p95 in the cytoplasm
`was also inhibited by TAPI, with little or no effect at 1 /XMand
`maximal
`inhibition at 10 piM(Fig. 5). In three separate experiments,
`1
`P.MTAPI
`inhibited ECD and p95 levels by <50%, and in all cases,
`maximum inhibition was achieved by 10 JU.MTAPI. No change in
`pl85HER-2/neu
`levels could be detected in cells treated with TAPI or
`when shedding was stimulated by TPA and chloroquine
`(Fig. 4),
`probably because protcolytic processing of pi85 is constitutive
`and
`limited, with ~20% converted into soluble HER-2/neu in 2 h (13).
`Detection of pl85 and p95 HER-2/neu in Breast Cancer Tissue.
`Tumor
`tissues were homogeni/.ed.
`fractionated,
`and examined
`for
`HER-2/neu proteins by Western analysis. The membrane-enriched
`but
`
`tissues
`from some tumor
`the soluble fraction (data not shown)
`not
`contained the full-length product, p 185, and the truncated p95HER-
`2/neu protein that comigrated with HER-2/neu
`proteins
`from the
`control 17-3-1 cells (Fig. 6). In addition, p95, along with pl85. was
`detected in two of eight ovarian cancer tissues (data not shown). Initial
`analyses of several breast cancer
`tissues revealed distinct expression
`patterns of p95 and p 185. One group had no detectable p 185 or p95
`(Fig. 6, samples 39 and 69). A second category of specimens
`ex
`pressed both pi85 and p95 (Fig. 6, samples 60, 53, 04, and 22). An
`additional group contained p 185 with relatively little or no p95 (Fig.
`6, samples 40, 58, 38, 57, 17, and 75). As observed in previous
`studies, some samples were pl85 rich (Fig. 6, samples 04, 22, 57, 17,
`75). The samples that were characterized as highly positive for pi85
`were initially identified by immunoassay
`values of >400 units (see
`"Materials
`and Methods"
`and Fig. 6 legend). The results of
`the
`
`Western analysis suggested that the tumors were heterogeneous with
`respect
`to HER-2/neu protein products and that they could be subdi
`vided based on the presence or absence of p95.
`Western analysis of 161 breast cancer samples revealed that 22.4%
`were p95 positive. The pl85-positivc
`samples were further subdivided
`into highly positive or HER-2-rich specimens, based on comparisons
`with HER-2/neu-overexpressing
`samples,
`identified by immunoassay
`and comparisons with the control
`17-3-1 extract,
`as described
`in
`"Materials
`and Methods." The highly positive p 185 samples
`repre
`sented 21.79!- of the total samples. All samples
`that expressed p95
`were also positive for p 185, although 65% of pl85-positive
`samples
`did not contain p95. Of the p95-positive
`samples, 63.9% were also
`highly positive for p 185, and 369!- had low p 185 levels.
`and
`Relationship between p95-positive, Highly p!85-positive,
`Other Prognostic Factors. The proportion of 78 node-positive breast
`cancer patients expressing
`p95 was higher
`than the proportion
`of
`node-negative
`patients expressing p95 (P = 0.032); pl85-rich
`sam
`ples had no significant association with node status (Table 1).
`Neither p95-positive nor pl85-rich samples correlated significantly
`
`Probe:
`
`anti-neu
`
`(C)
`
`Patient #
`
`(Conti
`
`60
`
`40
`
`58
`
`38
`
`53
`
`39
`
`B
`
`(coni)69 04 22 57 17 75
`
`Fig. 6. Western blotting analysis of 12 breast cancer tissues. Human intraductal breast
`cancer tissues were fractionated, and 20 ¿igof protein from 12 patients were subjected lo
`Western blotting with anti-neu(C)
`as described in "Materials and Methods." The control
`lane contained 3 /ng of protein from iransfected 3T3 cells, 17-3-1. The positions of p 185
`(¡ophand) and p95 ihoirotn hand) are marked in the control 17-3-1 sample in B. A
`photograph of the film that was exposed to the membrane for 20 min (A) and for 5 min
`(ß)is shown. HER-2/neu immunoassay values were: <100 units for samples 60, 39. and
`69: 389 units for sample 40: 258 units for sample 5«;302 units for sample 38; 200 units
`for sample 53: 2.000 units for sample 04: 10,000 unils for sample 22; 1000 units for
`sample 57: 550 units for sample 17: f)74 units for sample 75.
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on November 12, 2014. © 1998 American Association for Cancer
`Research.
`
`5126
`
`IMMUNOGEN 2048, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`long-
`though the mechanism for this stimulation was not examined,
`term exposure of cells to TPA has been found to enhance internal-
`ization of RTKs (50), whereas chloroquine, an agent that alters the pH
`p95
`pi 85 highly
`
`positive29.514.327.817.732.019.723.722.1P.032NSNSNS%positive24.422.222.221.529.717.920.323.2PNS*NSNSNS
`in endosomes and lysosomes, may inhibit complete proteolytic break
`down or alter RTK trafficking (51 ). Finally, p95 and ECD were both
`inhibited by addition of the hydroxamate
`compound, TAPI,
`to intact
`cells, and both were maximally inhibited by s 10 /AMTAPI (Fig. 5).
`The strong inhibition by TAPI indicates that most of the ECD and p95
`in BT474 cells were generated by a metalloprotease
`(46, 47) and that
`this class of inhibitors may be effective
`in controlling
`shedding in
`breast cancer patients. Although p95 and shedding were modulated
`under several different conditions,
`changes
`in the p 185 levels could
`not be detected. Unlike several
`transmembrane
`proteins that only shed
`when induced by TPA, proteolytic shedding of p 185 occurs continu
`ally at a low basal
`level (11, 12), with only —¿(cid:3)20%converted into
`soluble ECD in 2 h (13). The truncated cell protein of M, -95,000
`described here was somewhat
`larger than the expected A/r 75,000-
`80,000 for the cytoplasmic
`remnant of the M,
`-105,000-110,000
`ECD. p95 or the ECD might migrate anomalously in gels because the
`site of cleavage for ECD shedding is not known. Although our studies
`showed that the ECD and p95 are coordinately produced in culture by
`proteolytic activity that is sensitive to a metalloprotease
`inhibitor,
`it is
`not yet known whether p95 levels in breast
`tumors will be directly
`coupled to serum ECD. In some cases, ECD may be the product of an
`alternative
`transcript
`(14), or the metabolism of p95 may vary in
`different cells. Future studies aimed at testing cancer tissue and serum
`from the same patients will be required to evaluate whether
`serum
`ECD correlates with tissue p95 in vivo.
`in transfected
`A HER-2/neu product of the same size, Mr 95,000,
`3T3 cells, cultured breast carcinoma
`cells, breast cancer
`tissue, and
`ovarian cancer
`tissue suggests
`that a similar proteolytic processing
`event may occur in the different cells. However, p95 was not detected
`in all cells and tumor tissue that contain p 185. Two nontumorigenic
`breast epithelial cell lines had no detectable p95 (Fig. 2). In addition,
`the SKOV3 ovarian carcinoma cells, which overexpress p 185, had a
`disproportionately
`low amount of p95 (Fig. 4). These observations
`indicate that production of p95 is regulated. The cells with variable
`levels of truncated HER-2/neu products may differ in the amount of
`the relevant protease activity or the protein substrate may have an
`altered conformation affecting sensitivity to proteolytic cleavage.
`p95HER-2/neu
`has properties
`that suggested a rationale for exam
`ining its association with prognostic
`factors in breast cancers.
`It has
`kinase activity,
`is tyrosine-phosphorylated,
`suggesting its activity in
`signaling, and is truncated from its NH2 terminus. Oncogenic signal
`ing by HER-2/neu is known to depend on its level of kinase activity
`(33-35). Because p95 was at 100% of p 185 in some breast cancer
`samples,
`it may impact
`the amplitude of the kinase signal. Moreover,
`an NH2-terminally truncated kinase domain such as p95 is expected to
`emit a constitutive
`signal by analogy to results with engineered
`deletions of the ECD from the HER-2/neu product
`(32-36). Taken
`
`FactorNodesPositive
`
`(78)Negative
`(63)Tumor
`size(cm)5=3
`
`(54)<3
`(79)ERNegative
`
`(37)Positive
`(117)PRNegative
`
`(59)Positive
`(95)%
`"161 samples were examined by Western analysis. See "Materials and Methods"
`
`for
`
`of patient material used and methods of analysis. Not all samples had
`a description
`information for the factors examined.
`' NS. not significant.
`
`with other factors known to predict poor prognosis (49), including ER
`and PR negativity and tumor size of &3 cm (Table 1).
`Influence of p95 in the Highly pl85-positive Group. We ques
`tioned why a similar percentage of node-positive
`and node-negative
`patients were p 185 rich (24.4% versus 22.2%, Table 1), whereas p95
`was associated with node-positive
`patients
`(65.7% of the pl85-rich
`samples contained p95). We,
`therefore, examined whether
`the pres
`ence
`or absence
`of p95 in the
`specimens
`that overexpressed
`p 185HER-2/neu
`affected the relationship with lymph node status
`(Table 2). The highly pl85-positive
`samples
`that contained
`p95
`(n = 21) had a significantly higher association with metastasis to the
`lymph nodes, whereas
`the highly pl85-positive
`samples
`that were
`negative for p95 (n = 11) were associated with lymph nod